Cigarette smoking is a leading preventable cause of death, responsible for one in five deaths annually in the United States alone. Many adult smokers want to or try to quit smoking, but, are unable to do so due to lack of personalized cessation treatments that ultimately leads to successful quitting. The primary hurdle has been the unavailability of technology that can be used to collect continuous measurements of stress contexts in field on predictors of abstinence and relapse. Without these measurements, it is difficult to understand the psychophysiological and biobehavioral stress mechanisms that may predict smoke lapse. This application proposes to use a new field deployable tool called AutoSense that was developed under NIH's GEI program to collect personal psychophysiological measures of stress and mediating social and environmental contexts in the natural environment of individuals to discover signatures of early relapse. AutoSense consists of several wireless sensors (e.g., ECG, respiration, triaxial accelerometer) attached to a chest band that transmit sensor measurements wirelessly to a smart phone. Fine-grained measurement of physiology enables computation of physiological measures that are predictive of stress such as minute ventilation and heart rate variability. The project uses sophisticated machine learning methods to estimate stress levels, smoking episodes and conversation episodes from physiological measurements. In addition, accelerometers are used to identify changes in posture and physical activity, and GPS measurements on the mobile phone are used to identify location and transportation modality. Seventy-two smokers who want to quit will be recruited. Two weeks prior to quit date, they will wear AutoSense for 24h in the field. Behavioral treatment for abstinence will begin in their visit to the lab prior to quitting. They wil wear AutoSense for 72h in the field, beginning on the quit date. In addition to collecting continuous measurements across all sensors, smoking and lapse events will be self-reported. Saliva samples will be collected for cortisol assessment, in addition to physiological measures of stress. Participants will report back to the lab one week after the quit date to determine smoking/lapse status. The aim of this project is to identify psychophysiological measures of stress and environmental cues that may predict early lapse. First, it will discover patterns in physiological and hormonal measures in response to withdrawal stress and acute stress that may predict lapse in the first week. Second, it will investigate the role of social and environmental context such as conversation behavior (e.g., duration and frequency of conversations), lifestyle factors (e.g., time spent in seated posture and in commuting), location (time spent at home vs. office), and indoor/outdoor status in mediating the relationship between stress and early lapse.